## 1. Danish Exports that Do Not Cross Border

## Source details

**Canonical URL:** [1. Danish Exports that Do Not Cross Border](https://www.imf.org/-/media/files/publications/cr/2024/english/1dnkea2024002-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2024/english/1dnkea2024002-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2024/english/1dnkea2024002-print-pdf.pdf.json)

---

### A. Introduction
- The pharmaceutical industry in Denmark has grown rapidly: since 2005, the real gross value added (GVA) of pharmaceuticals has increased sixfold.
- The recent surge in real GVA reflects increased export demand for new weight-loss drugs developed by Novo Nordisk.
- Two structural characteristics:
  - Danish large multinational enterprises (MNEs) increasingly rely on “merchanting and processing (M&P),” with increased exports that never cross Danish borders. In pharmaceuticals, much of value is attributable to intellectual property and production commonly uses contract manufacturers abroad.
  - Expansion of the pharmaceutical sector has been mainly driven by a significant increase in labor productivity: labor productivity in the pharmaceutical sector has increased more than threefold since 2005, compared to a 20 percent increase for all industries.
- Analytical focus and data:
  - Quantitative relationship between productivity shocks at pharmaceutical companies (notably Novo Nordisk) and real GDP growth.
  - Microfoundations: whether productivity shocks in a dominant firm like Novo Nordisk spill over to other firms, within and outside the pharmaceutical industry.
  - Firm-level data from Orbis; regressions span 1993–2022 for GDP impact and 1995–2022 for spillover analysis.

### Box: Danish Exports that Do Not Cross Border (Key points)
- Exports from Danish companies’ production abroad are classified into merchanting and processing depending on ownership of intermediate inputs; such exports are counted as Danish exports in Balance of Payment Statistics although goods never cross Danish borders.
- Processing by large manufacturers (including pharmaceuticals) drives the trend increase in foreign production.
- Implications of increased foreign production:
  - Foreign production combines factor inputs from Denmark and the host country; Danish factor inputs are heavier in intangible assets and lighter in labor.
  - Foreign production exhibits higher labor productivity and lower employee compensation.
  - Aggregate labor productivity growth may not reflect domestic economy labor productivity growth; output gap based on aggregate output may overstate capacity pressures in the labor market.

### B. Danish Pharmaceutical Sector and the Macroeconomy (Stylized facts and magnitudes)
- Novo Nordisk:
  - Largest pharmaceutical company in Denmark; among top 20 largest pharmaceutical companies globally by revenue.
  - Sales as a share of Denmark’s GDP increased from 1 percent in the early 1990s to 8.3 percent in 2023.
  - To meet strong foreign demand for diabetes and obesity drugs, Novo Nordisk is expanding production capacity both in Denmark and abroad.
- Macro contributions of pharmaceutical industry (selected magnitudes):
  - Real GDP growth contributions:
    - Pharmaceutical sector contributed about 10 percent to real GDP growth during 2020–21.
    - Contributions surged to 90 percent in 2022 and 50 percent in 2023.
  - Exports:
    - Danish exports originating from Denmark have fluctuated around 29–33 percent of GDP over the past 15 years.
    - Pharmaceutical exports increased from around 2 percent of GDP to 6 percent of GDP since 2007.
    - Exports not crossing the border increased from about 1 percent of GDP to about 10 percent of GDP.
  - Tax revenue:
    - Corporate income tax payments by Novo Nordisk doubled over the last 5 years, increasing from about 0.3 percent of GDP in 2019 to 0.6 percent of GDP in 2023.
  - Employment:
    - Pharmaceutical industry accounts for 6.7 percent of nominal value added while its share in employment is around 1 percent.
    - The sector’s employment share remains relatively low due to reliance on knowledge and significant foreign production.

### C. Empirics: Methodology (Shocks construction and regression approach)
- ScaledShocks construction:
  1. Firm-level labor productivity growth = ln(Sales/Employees)t − ln(Sales/Employees)t−1.
  2. De-mean to obtain firm-specific labor productivity shocks (remove economy-wide effect).
  3. Scale firm-specific shocks by firm size measured by sales: ScaledShock_j,t = (Sales_j,t−1 / GDP_t−1) * firm-specific shock_j,t.
- Two shock aggregates:
  - Novo Nordisk-specific ScaledShocks.
  - ScaledShocks aggregated across the top-10 firms each year to capture large MNE shocks.
- Regressions:
  - GDP growth regressions follow Gabaix (2011) approach: regress log differences of real GDP on ScaledShocks and lags; estimation by OLS.
  - Spillover regressions use firm-fixed effects and controls including lagged firm-specific shocks, cash flows, leverage, loss, and size; interaction term with a pharmaceutical-industry dummy (P_dummy or Pharma_Dummy) measures within-industry spillovers.

### D. Empirics: Impact on Real GDP Growth (Key regression results)
- Table 2 summary (coefficients and t-statistics preserved):
  - Column (1): ScaledShocks(Novo)
    - ScaledShocks(Novo)t: -0.00795 (t = -0.53)
    - ScaledShocks(Novo)t-1: 0.0113 (t = 0.92)
    - ScaledShocks(Novo)t-2: 0.0236* (t = 1.75)  (* p<0.1)
    - Constant: 1.704*** (t = 3.61)
    - N = 25, R2 = 0.169
  - Column (2): ScaledShocks(Top 10)
    - ScaledShocks(Top 10)t: 0.00336** (t = 2.53)  (** p<0.05)
    - ScaledShocks(Top 10)t-1: -0.0000102 (t = -0.00)
    - ScaledShocks(Top 10)t-2: 0.00365* (t = 1.87)  (* p<0.1)
    - Constant: 1.819*** (t = 5.15)
    - N = 30, R2 = 0.201
- Interpretation and illustrative magnitude:
  - Weak evidence that Novo Nordisk’s productivity shock is positively associated with GDP growth (Column 1); only the second lag significant at the 10 percent level.
  - Stronger evidence that productivity shocks to the top-10 MNEs are positively associated with GDP growth (Column 2): contemporaneous coefficient significant at the 5 percent level and second lag at the 10 percent level.
  - Using Sales-to-GDP ratio of the top-10 MNEs for 2021, the estimated model suggests that a 1.85 percent increase in labor productivity was associated with a 0.3 percentage points increase in real GDP growth on average during 1993–2022.

### E. Empirics: Spillovers from Novo Nordisk to Other Firms (Specification and results)
- Spillover regression specification (summary):
  - ScaledShocks(Other)j,t = α_j + β1 ScaledShocks(Novo)i,t−1 + β2 ScaledShocks(Novo)i,t−1 * P_dummy + β3 X_j,t−1 + ε_j,t
  - Controls X include lagged firm-specific labor productivity shocks, cash flows, leverage, loss, and size; firm-fixed effects included.
- Table 3 / Table 4 summary (coefficients and t-statistics preserved):
  - Column (1): ScaledShocks(Novo)t-1 = 0.000339 (t = 1.11)
  - Column (2): ScaledShocks(Novo)t-1 = 0.000259 (t = 0.82); ScaledShocks(Novo)t-1 * P_Dummy = 0.00962** (t = 2.30) (** p<0.05)
  - Constant approx 1.022 and 1.009; N = 7,745 observations.
- Interpretation:
  - Correlations between Novo Nordisk’s productivity shocks and those of other firms are weak: estimated coefficient ~0.0003 and not statistically significant (Column 1).
  - Allowing for within-industry interaction, the direct spillover coefficient remains small and insignificant, but the interaction term (within-pharmaceutical-industry effect) is positive (0.00962) and statistically significant at the 5 percent level (Column 2).
  - Using company sizes in 2022, estimates imply:
    - One percent labor productivity shock to Novo Nordisk is associated with 0.06 percent of labor productivity shock to other firms next year (using 6.3 percent and 0.03 percent Sales-to-GDP ratios).
    - One percent of labor productivity shock to Novo is associated with 0.75 percent additional increase in labor productivity shock to other firms in the pharmaceutical industry next year (using 0.08 percent Sales-to-GDP ratio for pharmaceutical industry “other firms”).

### F. Key Findings and Takeaways
- Growth and externalization:
  - The pharmaceutical sector’s rapid growth has substantially raised real GVA and exports, driven largely by productivity gains and expanding foreign production.
  - A substantial share of pharmaceutical exports do not cross Denmark’s borders (merchanting and processing), implying that aggregate productivity and output measures may overstate domestic capacity pressures.
- Macroeconomic links:
  - Weak statistical evidence that Novo Nordisk’s firm-specific productivity shocks alone have large measurable effects on aggregate real GDP growth over the sample, though some positive association appears at lag.
  - Stronger evidence that cumulative shocks to the top-10 MNEs are positively associated with real GDP growth, with contemporaneous and lagged coefficients statistically significant.
  - Spillover analysis indicates limited and uncertain spillovers from Novo Nordisk to the broader set of firms but a significant within-industry spillover effect to other pharmaceutical firms.
- Policy-relevant implication:
  - Dominant multinational firms that rely on foreign production can concentrate effects in value-added and export statistics while generating more modest direct domestic employment effects; within-industry knowledge and productivity diffusion appears important.
- Additional concluding notes:
  - Limited risk that Denmark’s booming pharmaceutical company would become its “Nokia.”
  - Most pharmaceutical production occurs overseas under Danish ownership, limiting domestic linkages in employment and supply chains.
  - Empirical results may underestimate Novo Nordisk’s influence due to limited data.

### Key quantitative findings and interpretation of regression coefficients
- Top-10 MNEs to GDP and aggregate impact:
  - The regression coefficient in (2) in Table 2: 100 units increase in ScaledShocks(Top 10)t is associated with 0.34 percentage point increase in real GDP growth.
  - In 2021, sum of top-10 firms’ sales was 54 percent of GDP.
  - Assuming top-10 firms have same level of productivity shocks in 2021, 100 units increase in ScaledShocks(Top 10)t translates to 1.85 percent increase in labor productivity.
  - Therefore, 1.85 percent increase in labor productivity by top 10 firms is associated with 0.34 percentage point increase in real GDP growth.
- Novo Nordisk firm-specific spillovers:
  - Regression coefficient for Novo Nordisk firm-specific shock in (1) in Table 4: one unit increase in weighted Novo Nordisk labor productivity shock leads to 0.0003 unit increase in weighted “other firms” shock (statistical significance issue aside).
  - Using 6.3 percent Sales to GDP ratio for Novo Nordisk and 0.03 percent of sales to GDP ratio of “other firms” in the sample for 2022:
    - One percent labor productivity shock to Novo Nordisk is associated with 0.06 percent of labor productivity shock to other firms next year.
  - Regression coefficient for pharmaceutical industry interaction term in (2) in Table 4: one unit increase in weighted Novo labor productivity shock has 0.00962 unit additional increase in weighted pharmaceutical industry “other firms”.
    - Using 0.08 percent of Sales to GDP ratio for pharmaceutical industry “other firms”, one percent of labor productivity shock to Novo is associated with 0.75 percent additional increase in labor productivity shock to other firms in the pharmaceutical industry next year.

### NBFI sector vulnerabilities and risks (selected findings)
- Size and growth:
  - Denmark’s NBFI sector has tripled in size since the GFC.
  - The NBFI sector now accounts for almost 300 percent of GDP.
  - The share of NBFIs in total assets of the financial system increased from less than 30 percent in 2007 to almost 45 percent in 2023.
- Key vulnerabilities:
  - Asset price exposures: high exposure to debt securities and equities; significant allocation to covered bonds issued by mortgage credit institutions.
  - Leverage: increased but not excessive relative to the EA average; insurance companies’ debt-to-assets increased since the GFC but lower than EA average; pension and investment funds’ gross and adjusted leverage show increases slightly above the EA average.
  - Liquidity: liquidity buffers sizeable if covered bonds are included (Liquidity 1); Liquidity 1 above EA average for pension funds and at par for investment funds; insurance companies’ Liquidity 1 below EA average. Liquidity 2 (cash and deposits only) indicates lower buffers compared to the EA average, especially for investment funds.
  - Interconnectedness: high domestic and cross-border interconnectedness; large exposures to a few countries such as the U.S., the UK, Luxembourg, and Ireland.
- Risks:
  - Market risk: elevated but declining; potential for disorderly falls in asset prices.
  - Credit risk and macro-financial risk: Denmark may be subject to a macro-financial risk attributed to a sizeable wealth effect.
  - Historical stress episodes: GFC and the 2022 monetary policy tightening tested resilience, resulting in significant losses for NBFIs but the system proved resilient.

### Corporate credit risks, macro-financial risk, and mitigating factors
- Corporate credit risks:
  - Higher interest rates made it more challenging for some Danish corporates to service their debt, especially in industry, construction, trade, and real estate.
  - Corporate credit risk parameters indicate elevated credit risks (EBA, 2024).
- Macro-Financial Risk:
  - Danish households hold very large assets combined with a high level of household debt, making Denmark potentially prone to a significant wealth effect.
  - Under an adverse scenario involving losses in household wealth, Denmark is estimated to be markedly impacted in terms of private consumption.
  - Historical evidence: household consumption in Denmark dropped by more than 6 percent in 2008.
- Mitigating factors:
  - Pension companies hold about 50 percent of total liabilities in non-guaranteed market return products.
  - Significant exposure to real estate is primarily indirect through covered bonds; covered bonds demonstrated resilience during stress episodes.
  - NBFIs’ credit provision accounts for only around 10 percent of GDP.

### Policy recommendations and implications
- Strengthen systemic risk assessment and policy framework for NBFIs:
  - Develop a comprehensive systemic risk assessment framework covering both banks and NBFIs.
  - Develop a system-wide stress testing framework combining banking and nonbanking institutions.
- Close monitoring warranted on:
  - Leverage, liquidity buffers, and interconnectedness among NBFIs.
  - Asset price exposures, including indirect exposures via covered bonds and CRE vulnerabilities.
- Consider international experience and coordination given cross-border interconnectedness and common exposures.

### Box 1 — An Overview of the Danish Covered Bond Market (selected facts)
- Outstanding issuance: €436 billion (about 123 percent of GDP) as of end-2023.
- Currency composition: 96.8 percent denominated in Danish Krone (DKK); Euro (EUR) and Swedish Krona (SEK) each account for 1.6 percent.
- Asset pool safeguards: maximum loan-to-value limits at 80 percent residential and 60 percent commercial; overcollateralization exceeds ECBC’s minimum requirement of 2 percent.
- Historical stress episodes:
  - GFC in 2008: sudden liquidity dry-up prompting Danmarks Nationalbank intervention.
  - Covid pandemic in 2020: spread widening with rapid subsequent recovery.
- Investor base and holdings:
  - Domestic holdings about DKK 2.8 billion in March 2024.
  - Sectoral composition changes (2009 to 2024): banks decreased from 58 percent to 34 percent; insurance companies and pension funds increased from 21 percent to 35 percent; investment funds increased from 9 percent to 13 percent.
  - Foreign holdings rose from 9 percent in 2008 to 25 percent as of end-2023 (peak at 33 percent in 2020 and 2021); Euro Area countries constitute about 56 percent of foreign holdings.
- Policy implication: covered bond market remains stable with changing dynamics; vulnerabilities and risks should continue to be monitored, including continuous review of the macroprudential policy stance.

*Source — 1dnkea2024002-print-pdf*

### 1. Danish Exports that Do Not Cross Border ______________________________________________ 4

### PRODUCTIVITY SHOCKS TO THE PHARMACEUTICAL SECTOR AND THE DANISH ECONOMY

### A. Introduction
- The pharmaceutical industry in Denmark has grown rapidly in recent years; since 2005, the real gross value added (GVA) of pharmaceuticals has increased sixfold.
- The recent surge in real GVA reflects increased export demand for new weight-loss drugs developed by Novo Nordisk.
- Two structural characteristics highlighted:
  - Danish large multinational enterprises (MNEs) increasingly rely on “merchanting and processing (M&P),” with increased exports that never cross Danish borders. In pharmaceuticals, much of value is attributable to intellectual property and production commonly uses contract manufacturers abroad.
  - Expansion of the pharmaceutical sector has been mainly driven by a significant increase in labor productivity: labor productivity in the pharmaceutical sector has increased more than threefold since 2005, compared to a 20 percent increase for all industries.
- Analytical focus:
  - Quantitative relationship between productivity shocks at pharmaceutical companies (notably Novo Nordisk) and real GDP growth.
  - Microfoundations: whether productivity shocks in a dominant firm like Novo Nordisk spill over to other firms, within and outside the pharmaceutical industry.
- Data: firm-level data from Orbis; regressions span 1993–2022 for GDP impact and 1995–2022 for spillover analysis.

### Box: Danish Exports that Do Not Cross Border (Key points)
- Exports from Danish companies’ production abroad are classified into merchanting and processing depending on ownership of intermediate inputs; such exports are counted as Danish exports in Balance of Payment Statistics although goods never cross Danish borders.
- Processing by large manufacturers (including pharmaceuticals) drives the trend increase in foreign production.
- Implications of increased foreign production:
  - Foreign production combines factor inputs from Denmark and the host country; Danish factor inputs are heavier in intangible assets and lighter in labor.
  - Foreign production exhibits higher labor productivity and lower employee compensation.
  - Aggregate labor productivity growth may not reflect domestic economy labor productivity growth; output gap based on aggregate output may overstate capacity pressures in the labor market.

### B. Danish Pharmaceutical Sector and the Macroeconomy (Stylized facts and magnitudes)
- Novo Nordisk:
  - Largest pharmaceutical company in Denmark; among top 20 largest pharmaceutical companies globally by revenue.
  - Sales as a share of Denmark’s GDP increased from 1 percent in the early 1990s to 8.3 percent in 2023.
  - To meet strong foreign demand for diabetes and obesity drugs, Novo Nordisk is expanding production capacity both in Denmark and abroad.
- Macro contributions of pharmaceutical industry (selected magnitudes):
  - Real GDP growth contributions:
    - Pharmaceutical sector contributed about 10 percent to real GDP growth during 2020–21.
    - Contributions surged to 90 percent in 2022 and 50 percent in 2023.
  - Exports:
    - Danish exports originating from Denmark have fluctuated around 29–33 percent of GDP over the past 15 years.
    - Pharmaceutical exports increased from around 2 percent of GDP to 6 percent of GDP since 2007.
    - Exports not crossing the border increased from about 1 percent of GDP to about 10 percent of GDP.
  - Tax revenue:
    - Corporate income tax payments by Novo Nordisk doubled over the last 5 years, increasing from about 0.3 percent of GDP in 2019 to 0.6 percent of GDP in 2023.
  - Employment:
    - Pharmaceutical industry accounts for 6.7 percent of nominal value added while its share in employment is around 1 percent.
    - The sector’s employment share remains relatively low due to reliance on knowledge and significant foreign production.

### C. Empirics: Shocks to Novo Nordisk and MNEs and their Impact on the Danish Economy (Methodology)
- Two sets of regressions:
  - Impact of labor productivity shocks at Novo Nordisk and other large MNEs on real GDP growth (1993–2022).
  - Spillover effects of labor productivity from Novo Nordisk to other firms within Denmark (1995–2022).
- Construction of ScaledShocks:
  1. Firm-level labor productivity growth = ln(Sales/Employees)t − ln(Sales/Employees)t−1.
  2. De-mean to obtain firm-specific labor productivity shocks (remove economy-wide effect).
  3. Scale firm-specific shocks by firm size measured by sales: ScaledShock_j,t = (Sales_j,t−1 / GDP_t−1) * firm-specific shock_j,t.
- Two shock aggregates:
  - Novo Nordisk-specific ScaledShocks.
  - ScaledShocks aggregated across the top-10 firms each year to capture large MNE shocks.
- Regression for GDP growth follows Gabaix (2011) approach: regress log differences of real GDP on ScaledShocks and lags; estimation by OLS.

### D. Empirical Results: Impact on Real GDP Growth (Key regression results)
- Table 2 summary (coefficients and significance preserved):
  - Column (1): ScaledShocks(Novo)
    - ScaledShocks(Novo)t: -0.00795 (t = -0.53)
    - ScaledShocks(Novo)t-1: 0.0113 (t = 0.92)
    - ScaledShocks(Novo)t-2: 0.0236* (t = 1.75)  (* p<0.1)
    - Constant: 1.704*** (t = 3.61)
    - N = 25, R2 = 0.169
  - Column (2): ScaledShocks(Top 10)
    - ScaledShocks(Top 10)t: 0.00336** (t = 2.53)  (** p<0.05)
    - ScaledShocks(Top 10)t-1: -0.0000102 (t = -0.00)
    - ScaledShocks(Top 10)t-2: 0.00365* (t = 1.87)  (* p<0.1)
    - Constant: 1.819*** (t = 5.15)
    - N = 30, R2 = 0.201
- Interpretation and illustrative magnitude:
  - Weak evidence that Novo Nordisk’s productivity shock is positively associated with GDP growth (Column 1); only the second lag significant at the 10 percent level.
  - Stronger evidence that productivity shocks to the top-10 MNEs are positively associated with GDP growth (Column 2): contemporaneous coefficient significant at the 5 percent level and second lag at the 10 percent level.
  - Using Sales-to-GDP ratio of the top-10 MNEs for 2021, the estimated model suggests that a 1.85 percent increase in labor productivity was associated with a 0.3 percentage points increase in real GDP growth on average during 1993–2022.

### E. Empirics: Spillovers from Novo Nordisk to Other Firms (Specification and results)
- Spillover regression specification:
  - ScaledShocks(Other)j,t = α_j + β1 ScaledShocks(Novo)i,t−1 + β2 ScaledShocks(Novo)i,t−1 * P_dummy + β3 X_j,t−1 + ε_j,t
  - ScaledShocks(Other) refers to size-weighted firm-specific labor productivity shocks for non-top-10 companies.
  - Control variables X include firms’ lagged firm-specific labor productivity shocks, cash flows, leverage, loss, and size; firm-fixed effects included.
  - Interaction term ScaledShocks(Novo) * P_dummy measures additional within-industry (pharmaceutical) spillovers.
- Table 3 summary (coefficients and significance preserved):
  - Column (1): ScaledShocks(Novo)t-1 = 0.000339 (t = 1.11)
  - Column (2): ScaledShocks(Novo)t-1 = 0.000259 (t = 0.82); ScaledShocks(Novo)t-1 * P_Dummy = 0.00962** (t = 2.30) (** p<0.05)
  - Constant approx 1.022 and 1.009; N = 7,745 observations.
- Interpretation:
  - Correlations between Novo Nordisk’s productivity shocks and those of other firms are weak: estimated coefficient ~0.0003 and not statistically significant (Column 1).
  - When allowing for within-industry interaction, the direct spillover coefficient remains small and insignificant, but the interaction term (within-pharmaceutical-industry effect) is positive (0.00962) and statistically significant at the 5 percent level (Column 2).
  - Using company sizes in 2022, estimates imply that on average a one percentage point increase in labor productivity of Novo Nordisk was associated with a (text truncated in source) — empirical evidence points to a significant within-industry spillover but limited broader-economy spillovers.

### F. Key Findings and Takeaways
- The pharmaceutical sector’s rapid growth has substantially raised real GVA and exports, driven largely by productivity gains and expanding foreign production.
- A substantial share of pharmaceutical exports do not cross Denmark’s borders (merchanting and processing), implying that aggregate productivity and output measures may overstate domestic capacity pressures.
- Empirical results:
  - Weak statistical evidence that Novo Nordisk’s firm-specific productivity shocks alone have large measurable effects on aggregate real GDP growth over the sample, though some positive association appears at lag.
  - Stronger evidence that cumulative shocks to the top-10 MNEs are positively associated with real GDP growth, with contemporaneous and lagged coefficients statistically significant.
  - Spillover analysis indicates limited and uncertain spillovers from Novo Nordisk to the broader set of firms but a significant within-industry spillover effect to other pharmaceutical firms.
- Policy-relevant implication:
  - The economic impact of dominant multinational firms that rely on foreign production can be concentrated in value-added and export statistics while generating more modest direct domestic employment effects; within-industry knowledge and productivity diffusion appears important.

*Prepared by Takuji Komatsuzaki with support from Fuda Jiang (all EUR).*

### 0.75 percentage point increase in labor productivity of other pharma companies. The strong

### 0.75 percentage point increase in labor productivity of other pharma companies. The strong

### Conclusions
- The pharmaceutical industry has substantial and increasing influence on the Danish economy:
  - Its contribution to real GDP growth has been extraordinary over the past two years.
  - It significantly contributes to exports.
  - Employment in the pharmaceutical sector remains low.
  - Contributions to tax revenues have increased from low levels.
- Empirical evidence on spillovers and macro linkages is mixed:
  - Productivity shocks at Novo Nordisk and Denmark’s overall economic growth appear to be positively correlated, albeit with lags, but the evidence is relatively weak.
  - Stronger evidence of positive correlations exists between the top-10 MNEs and Denmark’s GDP.
  - The spillover effect of Novo Nordisk’s productivity on the broader economy is small and uncertain.
  - There is a strong within-industry spillover effect in the pharmaceutical sector.
- Implications:
  - Limited risk that Denmark’s booming pharmaceutical company would become its “Nokia.”
  - Most pharmaceutical production occurs overseas under Danish ownership, limiting domestic linkages in employment and supply chains.
  - Empirical results may underestimate Novo Nordisk’s influence due to limited data.

### Key quantitative findings and interpretation of regression coefficients
- Top-10 MNEs to GDP and aggregate impact:
  - The regression coefficient in (2) in Table 2: 100 units increase in ScaledShocks(Top 10)t is associated with 0.34 percentage point increase in real GDP growth.
  - In 2021, sum of top-10 firms’ sales was 54 percent of GDP.
  - Assuming top-10 firms have same level of productivity shocks in 2021, 100 units increase in ScaledShocks(Top 10)t translates to 1.85 percent increase in labor productivity.
  - Therefore, 1.85 percent increase in labor productivity by top 10 firms is associated with 0.34 percentage point increase in real GDP growth.
- Novo Nordisk firm-specific spillovers:
  - Regression coefficient for Novo Nordisk firm-specific shock in (1) in Table 4: one unit increase in weighted Novo Nordisk labor productivity shock leads to 0.0003 unit increase in weighted “other firms” shock (statistical significance issue aside).
  - Using 6.3 percent Sales to GDP ratio for Novo Nordisk and 0.03 percent of sales to GDP ratio of “other firms” in the sample for 2022:
    - One percent labor productivity shock to Novo Nordisk is associated with 0.06 percent of labor productivity shock to other firms next year.
  - Regression coefficient for pharmaceutical industry interaction term in (2) in Table 4: one unit increase in weighted Novo labor productivity shock has 0.00962 unit additional increase in weighted pharmaceutical industry “other firms”.
    - Using 0.08 percent of Sales to GDP ratio for pharmaceutical industry “other firms”, one percent of labor productivity shock to Novo is associated with 0.75 percent additional increase in labor productivity shock to other firms in the pharmaceutical industry next year.
- Selected coefficients and statistics from Annex I. Table 2 (Regression Results: Impact of Firm-Specific Shocks to the Smaller Firms):
  - ScaledShocks(Novo)t-1: 0.000339 (t = 1.11) in column (1); 0.000259 (t = 0.82) in column (2).
  - ScaledShocks(Other)t-1: -0.295*** (t = -4.53) in column (1); -0.297*** (t = -4.59) in column (2). Note: * p<0.1, ** p<0.05, *** p<0.01.
  - Cash_flow_asset t-1: -0.349 (t = -1.04) in column (1); -0.337 (t = -1.02) in column (2).
  - Leverage t-1: 0.268** (t = 2.30) in column (1); 0.256** (t = 2.14) in column (2).
  - Loss t-1: -0.0501 (t = -0.76) in column (1); -0.0462 (t = -0.70) in column (2).
  - Size t-1: -0.0720 (t = -0.97) in column (1); -0.0709 (t = -0.95) in column (2).
  - ScaledShocks(Novo)t-1 * Pharma_Dummy: 0.00962** (t = 2.30) in column (2).
  - Constant: 1.022 (t = 0.81) in column (1); 1.009 (t = 0.80) in column (2).
  - N = 7745.

### Analysis of NBFI sector vulnerabilities and risks (selected findings)
- Size and growth:
  - Denmark’s NBFI sector has tripled in size since the GFC.
  - The NBFI sector now accounts for almost 300 percent of GDP.
  - The share of NBFIs in total assets of the financial system increased from less than 30 percent in 2007 to almost 45 percent in 2023.
- Key vulnerabilities identified: asset prices, leverage, liquidity, and interconnectedness.
  - Asset price exposures:
    - High exposure to debt securities and equities increases susceptibility to market risk via marked-to-market losses, volatility, incomplete hedging, and misalignment in collateral values.
    - Significant asset allocation to covered bonds issued by mortgage credit institutions indirectly exposes NBFIs to fluctuations in real estate markets.
  - Leverage:
    - Leverage has increased but does not seem excessive relative to the EA average.
    - For insurance companies, a general measure of leverage (debt to assets) increased since the GFC but tends to be lower compared to the EA average.
    - For pension and investment funds, gross leverage and adjusted leverage (excluding derivatives) show increases slightly above the EA average in these segments.
  - Liquidity:
    - Liquidity buffers are sizeable if covered bonds are included (liquidity 1).
    - Liquidity 1 indicates above EA average for pension funds and at par for investment funds; insurance companies’ liquidity buffer is below the EA average.
    - Liquidity 2 (cash and deposits only) indicates lower buffers compared to the EA average, especially for investment funds.
  - Interconnectedness:
    - High degree of domestic and cross-border interconnectedness.
    - Domestic: insurers and pension funds have significant interdependence with investment funds; NBFIs hold covered bonds more than EA average, increasing susceptibility to real estate markets and market pressures.
    - Cross-border: substantial cross-border asset allocation in equities and debt securities; large exposures to a few countries such as the U.S., the UK, Luxembourg, and Ireland.
- Risks:
  - Market risk: elevated but declining market risk for global and European markets; potential for disorderly falls in asset prices.
  - Credit risk and macro-financial risk: Denmark may be subject to a macro-financial risk attributed to a sizeable wealth effect.
  - Historical stress episodes: GFC and the 2022 monetary policy tightening tested resilience, resulting in significant losses for NBFIs but the system proved resilient.

### Policy recommendations and implications
- Strengthen systemic risk assessment and policy framework for NBFIs:
  - Develop a systemic risk assessment framework covering both banks and NBFIs.
  - Develop an ensuing system-wide stress testing framework.
- Close monitoring warranted on specific dimensions:
  - Leverage, liquidity buffers, and interconnectedness among NBFIs.
  - Asset price exposures, including indirect exposures via covered bonds and CRE vulnerabilities.
- Consider international experience and coordination given cross-border interconnectedness and common exposures.

*Italic: Source — 1dnkea2024002-print-pdf*

### 11.      Corporate credit risks remain. The DN (2024) reports that higher interest rates made it

### 11. Corporate credit risks remain.

### Corporate credit risks
- Higher interest rates made it more challenging for some Danish corporates to service their debt with earnings, especially in sectors such as industry, construction, trade, and real estate.
- Corporate credit risk parameters, including probability of default and loss given default, confirm elevated credit risks (EBA, 2024).
- While equity prices have continued to rise over the past couple of years, a deterioration in economic circumstances in Denmark and key advanced economies could put downward pressure on equity markets, adversely affecting the value of equities held by Danish NBFIs.

### Macro-Financial Risk
- Danish households hold very large assets combined with a high level of household debt (IMF, 2018), making Denmark potentially prone to a significant wealth effect.
- Under an adverse scenario involving losses in household wealth, Denmark is estimated to be markedly impacted in terms of private consumption, reflecting a strong wealth effect (Hviid and Kuchler, 2017).
- The wealth effect in Denmark is stronger than in most other advanced economies, especially when combining financial asset and housing wealth effects (Slacalek, 2009).
- Historical evidence: household consumption in Denmark dropped by more than 6 percent in 2008 (Andersen and others, 2016).

### Mitigating Factors
- Pension companies hold one of the highest shares of non-guaranteed market return products in Europe; these products account for about 50 percent of total liabilities, significantly mitigating market and credit risk impacts on solvency and liquidity.
- Significant exposure to real estate markets is primarily indirect, mainly through covered bonds, which have demonstrated considerable resilience even during stress episodes; this resilience supports liquidity risk management of investment funds given their sizable holding of covered bonds.
- NBFIs play a limited role in the credit market, with their credit provision accounting for only around 10 percent of GDP.

### Policy Options
- The authorities could consider strengthening their systemic risk assessment and policy framework for NBFIs by:
  - (i) developing a comprehensive systemic risk assessment framework covering both banking and nonbanking institutions; and
  - (ii) developing a system-wide stress testing framework combining banking and nonbanking institutions.

### Systemic Risk Assessment
- Implementing a comprehensive systemic risk assessment framework that covers both banks and NBFIs is crucial for countries with sizeable and interconnected financial systems.
- Examples of expanded coverage in other jurisdictions and institutions:
  - Financial stability reviews now routinely include systemic risk assessments for NBFIs (Central Bank of Ireland, 2024; Bank of England, 2024; Bank of Canada, 2024).
  - Dutch Central Bank (2024) and the Riksbank (2024) explicitly and extensively discuss emerging risks associated with NBFIs in jurisdictions where the NBFI sector is as sizable as in Denmark.
  - The FSB (2021) has built a framework relying on a set of indicators covering asset prices, leverage, liquidity, and domestic and cross-border interconnectedness.
  - The Central Bank of Ireland (2023) publishes a comprehensive heatmap gauging systemic risk arising from NBFIs as part of the Systemic Risk Pack at least once a year.
- EU-level and sectoral initiatives (summarized in the source) include proposals and ongoing policy initiatives to develop a macroprudential policy toolkit for NBFIs, and specific measures for insurance companies, pension funds, and investment funds.

### Stress Testing
- Developing a system-wide stress testing framework is an integral element of systemic risk assessment embedding NBFIs.
- International guidance and recent practices:
  - ESMA (2019, 2020) provides guidelines for developing a stress-testing liquidity framework for investment funds.
  - The ECB (2024) has started making explicit quantitative assessments of risks associated with NBFIs in their systemic risk assessments.
  - The BoE (2023) launched a stress test combining banks and NBFIs (including insurers, pension funds, investment funds, and central counterparties) to enhance understanding of risks posed by and to NBFIs, analyze NBFI and bank behaviors in stress, and investigate amplification channels to financial stability; efforts are carried out in collaboration with microprudential supervisors.

### Conclusions and Policy Considerations
- Systemic risk associated with NBFIs in Denmark appears to have been contained but requires close monitoring.
- The NBFI sector in Denmark has substantially increased in size since the GFC, bringing benefits but also risks; supervisors need to closely monitor leverage, liquidity buffers, and domestic and cross-border interconnectedness.
- Mitigating factors include a high share of non-guaranteed market-return products and NBFIs’ limited direct exposure to real estate markets and credit extension.
- Strengthening systemic risk assessment and the policy framework for NBFIs is warranted; developing a comprehensive framework covering both banks and NBFIs would pave the way for a system-wide stress testing framework combining banking and nonbanking institutions.

### Box 1 — An Overview of the Danish Covered Bond Market
- Outstanding issuance: €436 billion (about 123 percent of GDP) as of end-2023.
- Currency composition: 96.8 percent denominated in Danish Krone (DKK); Euro (EUR) and Swedish Krona (SEK) each account for 1.6 percent.
- Market instruments: callable bonds, bullet bonds, and floaters with or without a cap; market value concentrated in a few large series, majority issued by three large banks.
- Danish covered bond legislation emphasizes match funding and stringent asset liability management risk standards; pass-through products eliminate credit and market risk for mortgage banks and allow borrowers to terminate loans by buying back mortgage bonds (delivery option/buyback option).
- Credit ratings: generally stable, with over 90 percent of issuers maintaining a stable or positive rating outlook; covered bonds considered safe assets and have never defaulted, though their safety is linked to issuer financial health.
- Asset pool and safeguards: includes residential and commercial real estate with maximum loan-to-value limits at 80 percent and 60 percent, respectively; overcollateralization exceeds ECBC’s minimum requirement of 2 percent; no official limit on commercial real estate assets in the asset pool.
- Historical stress episodes:
  - During the GFC in 2008, the Danish covered bond market experienced a sudden liquidity dry-up, prompting Danmarks Nationalbank intervention.
  - At the onset of the Covid pandemic in 2020, covered bonds experienced spread widening with rapid subsequent recovery.
- Investor base changes since the GFC:
  - Domestic investors’ covered bond holdings increased to GFC levels in March 2024 at about DKK 2.8 billion.
  - Sectoral composition changes (2009 to 2024): banks (MFI) decreased from 58 percent to 34 percent; insurance companies and pension funds (ICPF) increased from 21 percent to 35 percent; investment funds increased from 9 percent to 13 percent.
  - Foreign holdings rose from 9 percent in 2008 to 25 percent of total outstanding Danish covered bonds as of end-2023 (peak at 33 percent in 2020 and 2021); Euro Area countries constitute about 56 percent of foreign holdings, with largest investments from Germany, Luxembourg, Ireland, Italy and Finland as of end-2023; Asian countries, in particular Japan, have also increased holdings.
  - Currency mismatch risk is low given about 96 percent of covered bonds are denominated in DKK, implying foreign holdings are mostly DKK-denominated.
- Policy implication: Denmark’s covered bond market remains stable with changing dynamics; vulnerabilities and risks should continue to be monitored, including continuous review of the macroprudential policy stance to safeguard the financial system.

*Source: IMF staff summary of the Denmark chapter.*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2024/english/1dnkea2024002-print-pdf.pdf_